Immuno-oncology was a super-hot topic in 2016. Several thousand publications contributed to the scientific progress, so choosing the top 3 is a challenging undertaking. But there are a few common themes seen in many highly ranked publications: the use of the latest hard- and software to gain an improved mechanistic understanding, why immuno-oncology actually works, and guidance on how to both design and predict individual patient success of novel therapies.
The study of interactions between various immune cell populations and genetically diverse cancer cell populations requires the development of novel tools which are capable of visualizing and analyzing protein expression levels of individual cells in their spatial context. Those tools, such as imaging CyTof, contribute significantly to the understanding of the disease, its progression and its response to therapy. The mechanistic understanding of the underlying systems biology will spin off new combination therapies and associated diagnostics. The profiling of cancer patients with respect to the mutational status of their tumor, and the status of their personal immune system, will enable physicians to identify the therapy which provides maximal value to that patient. As Bernd Bodenmiller states: “Exciting times are ahead”.
Multiplexed Epitope-Based Tissue Imaging for Discovery and Healthcare Applications
The review paper by Bernd Bodenmiller compares two recent technology developments that both simultaneously visualize a large number of different cell populations by multiplexed epitope-based imaging. The first method labels antibodies with fluorescent dyes, and the second method with metals. Bodenmiller discusses the pros and cons of both approaches, and he thinks beyond basic research towards the application in the clinical routine. Since the human optical system is optimized for three spectral channels (red, green, blue), and cannot handle more colors outside this spectrum, both approaches benefit tremendously by automated hyperspectral image analysis methods as it is described in .
Labelling with fluorescent dyes enables a spatial resolution as low as 200nm with typically five different antibodies and associated dyes per section. To overcome that limit, the dyes may be photobleached after image acquisition, and the tissue re-stained with another five antibodies. Although this procedure can be repeated up to 20 times, it is prone to autofluorescence errors and involves a significant amount of human interaction. The main advantage of this technique is that it uses common microscopes or tissue slide scanners, and can be therefore integrated into standard workflows.
Labelling with metals and imaging with a CyTOF instrument delivers a spatial resolution of only 1um with acquisition times up to 8 hours per tissue section. The main advantage of this method is the use of 32 labels simultaneously which enables even deeper profiling of T-cells, B-cells, NK-cells, macrophages, fibroblasts, tumor and endothelial cells. Moreover, various tumor subpopulations such as HER2(+), Estrogen Receptor(+) and Progesterone Receptor(+) cell populations may be shown. Neighborhood analysis of all those cells will very likely reveal novel biology and subsequently novel therapeutic options.
Integrative Analyses of Colorectal Cancer Show Immunoscore Is a Stronger Predictor of Patient Survival Than Microsatellite Instability
Bernhard Mlecnik et al investigated the mechanistic relationship of microsatellite instable (MSI) colorectal cancer and its immune contexture. The team discovered that MSI tumors entail higher densities of cytotoxic T cells (CD8 and GZMB), stromal cytotoxic T cells (CD8stroma), intratumoral cytotoxic T cells (CD8intra), B cells (CD20), macrophages (CD68 intra), and NK cells (NKp46) in the invasive margin compared to microsatellite stable (MSS) tumors. Furthermore, MSI tumors are associated with a higher Immunoscore, determined by measuring a distinct ratio of cytotoxic T (CD8) and memory T (CD45RO) cell densities in the tumor center and the defined invasive margin. By correlating MSI status and various immune scores derived from immune cell density measurements with overall and progression free survival, it turned out that the Immunoscore predicts patient outcome best.
This large scale study integrated heterogeneous data from multiple experimental setups such as brightfield and fluorescent immunohistochemistry, RNA sequencing, flow cytometry (FACS), and mutational analysis by fluorescent multiplex PCRs. Advanced bioinformatics analyses enabled the combination and cross-correlations of these data sets to revealed a novel holistic view on the underlying mechanisms of disease progression. Among other findings, the authors discovered that MSI tumors had more mutations of ACVR2A, TGFBR2, FBXW7, and ARID1A, and less mutations in APC, TP53, and KRAS. In particular, a mutated TGFBR2 class I neo-epitope could be associated with tumor-specific T cell response.
The publication concludes with specific recommendations for immune therapy of colon cancer: “(1) that MSI patients at an early stage might benefit the most from checkpoint T cell therapies, given that they have a strong effector T cell response and more frequently present a high Immunoscore associated with augmented PD1 expression, (2) that only a subgroup of metastatic MSI patients might benefit from checkpoint T cell therapies, namely those having a high Immunoscore, and (3) that MSS patients with tumors highly infiltrated with immune cells could also benefit from the therapy”.
Mechanism-driven biomarkers to guide immune checkpoint blockade in cancer therapy
Although checkpoint-based immune therapies for cancer, such as treatment with anti-CTLA4, anti-PD1 and anti-PDL1, celebrate response rates of up to 40% in melanoma and 20% in non-small cell lung cancer, the majority of patients do not benefit. In this remarkable review paper, Suzanne Topalian et al discuss various options to develop novel biomarkers using a mechanistic understanding of the tumor biology and its interaction with the host immune system. This holistic understanding is particularly needed with the clinical use of additional checkpoint inhibitors such as LAG3, TIM3, B7H3, CD39 and CD73, the double or even triple combinations of such antibodies, or even more combinations with “classical” targeted therapies against HER2 or EGFR mutations.
As a first example to develop immunological biomarkers to target the PD1/PDL1 pathways, the authors discuss the interaction of antigen presenting cells (APCs), CD4(+) T cells and CD8(+) T-cells in the lymphoid tissue, versus to the interactions of TH1 cells, Tregs, CD8(+) T-cells, macrophages and tumor cells in the tissue microenvironment. One major component of these interactions is interferon gamma signaling which indeed turned out to be a predictive factor for anti-PDL1 therapy. Other markers beyond PDL1 protein expression measured by immunohistochemistry involve more complex, multifactorial markers such as the close spatial proximity of PD1(+) to PDL1(+) cells, CD8(+)KI67(+) activated T cells, and the very recently discovered joint density of PDL1(+) and CD8(+) cells .
The second topic discussed in the paper are genetic biomarkers. The success of immune therapies has been clouded by the role of mutations of “driver oncogenes” in cancer cells. Some tumors drive PDL1 expression by oncogenetic signaling, such as silencing of PTEN or STAT3 activation. Strangely none of those oncogenes have been associated with response to anti-PDL1 therapy. This may however be attributed to missing data, in particular to multiplexed and spatially resolved mutational data together with immune cell profiling.
The analysis of pathways within cancer cells as well the analysis of the signaling network between cancer cells and immune cells will provide the necessary insight to progress science in the field of immune oncology.
 Sonja Althammer et al: Combinatorial CD8+ and PD-L1+ cell densities correlate with response and improved survival in non-small cell lung cancer (NSCLC) patients treated with durvalumab. DOI 10.1186/s40425-016-0191-4